A brain disease prevention information push system

Abstract

The present invention provides a brain disease prevention information push system, which relates to the field of medical health information technology. The system includes: a motion detection device, a communication component, a processor, a cloud server and an information receiving terminal. The processor is used to: obtain the relative position relationship between multiple motion detection devices at multiple moments within a preset time period; determine the posture data of the person to be tested according to the relative position relationship; and send the posture data to the cloud server through the communication component. The cloud server is used to: receive the posture data of the person to be tested; retrieve the brain CT image and electroencephalogram data of the person to be tested from the memory; obtain the brain disease prevention information of the person to be tested according to the posture data, brain CT image and electroencephalogram data; and push the brain disease prevention information to the information receiving terminal. According to the present invention, early brain diseases can be discovered in time and pushed to the information receiving terminal, prompting the person to be tested to pay attention to brain diseases.

Description

Technical Field

[0001] The present invention relates to the field of medical health information technology, and in particular to a brain disease prevention information push system. Background Art

[0002] CN117976124B discloses a disease prevention information push system and push method. A disease information push system includes a hospital information acquisition unit, a cloud server, and a terminal device; the cloud server is respectively connected to the hospital information acquisition unit and the terminal device by signal; the terminal device obtains the user's identity information and interest information, processes the interest information into first weight information, and the terminal device sends the first weight information and identity information to the cloud server. In this technical solution, when the cloud server pushes disease prevention information to the user, it will provide corresponding content after comprehensive judgment based on the user's first weight information and second weight information, so it is easier for the pushed disease prevention information to arouse the user's interest, so it can increase user stickiness and avoid wasting server resources.

[0003] CN105678066B discloses a disease self-diagnosis method for completing data training based on user feedback information, including the following steps: establishing a disease self-diagnosis data model, including disease data, disease type data, symptom data, and disease-symptom association data; the user obtains self-diagnosis results and treatment and medical advice through the disease self-diagnosis data model according to their own symptoms; the user gives feedback based on the self-diagnosis results or treatment and medical results; the disease self-diagnosis data model analyzes and learns based on the feedback, and continuously performs data training. The beneficial effects of this scheme are: through this scheme, a disease self-diagnosis data model is established, and a method for completing data training based on user feedback information is established, providing a complete data training approach and method for disease self-diagnosis information, and providing a more scientific and accurate basis for patients' self-diagnosis.

[0004] Therefore, although the relevant technology can enable the data collected on the user side to interact with the cloud server and perform self-diagnosis of diseases based on disease data, it is difficult to conduct comprehensive monitoring of specific brain diseases and it is impossible to push brain disease prevention information to prompt users to pay attention to brain diseases.

[0005] The information disclosed in the background technology section of this application is only intended to deepen the understanding of the general background technology of this application, and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0006] The present invention provides a brain disease prevention information push system, which can solve the technical problems that related technologies are difficult to comprehensively monitor specific brain diseases and difficult to increase users' attention to brain diseases.

[0007] According to the present invention, a brain disease prevention information push system is provided, comprising:

[0008] Motion detection equipment, communication components, processors, cloud servers and information receiving terminals;

[0009] The motion detection device includes a waist wearable motion detection device, a knee wearable motion detection device, and an ankle wearable motion detection device;

[0010] The processor is used to:

[0011] When the person to be tested wears the motion detection device, the relative position relationship between the multiple motion detection devices is obtained at multiple times within a preset time period;

[0012] Determining the posture data of the person to be measured according to the relative position relationship;

[0013] Sending the posture data at multiple times within a preset time period to a cloud server through a communication component;

[0014] The cloud server is used for:

[0015] Receiving posture data of the person to be tested;

[0016] Retrieving the brain CT image and EEG data of the person to be tested from the memory;

[0017] Obtaining brain disease prevention information of the person to be tested according to the posture data, the brain CT image and the EEG data;

[0018] The brain disease prevention information is pushed to an information receiving terminal.

[0019] Technical effect: According to the present invention, the posture data of the person to be tested can be monitored in real time by wearing a motion detection device, providing a data basis for timely detection of gait abnormalities, and can generate brain disease prevention information together with CT images, EEG data and other information, which can timely detect early brain diseases and push them to the information receiving terminal to prompt the person to be tested to pay attention to brain diseases. When training the posture similarity recognition model, the judgment accuracy of the posture similarity recognition model can be trained by the posture data in the historical database, and when the judgment is wrong, the wrong discriminant vector is used as a loss function, thereby improving the training intensity and reducing the probability of judgment errors. The posture data of control persons without brain diseases can also be used to train the posture similarity recognition model to eliminate the interference of non-disease factors, thereby improving the anti-interference ability and judgment accuracy of the posture similarity recognition model. Furthermore, the robustness of the posture similarity recognition model can be trained by the posture data of experimental patients with brain diseases who do not belong to the historical database, so that the posture similarity recognition model can be applied to the discrimination of posture data outside the historical database, thereby improving the performance of the posture similarity recognition model. When determining the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease, the first critical point can be screened by the gradient change of the pixel values ​​of the pixel points on the first circular test line, and the average width of the sylvian fissure can be determined by the conditional function, and then the average width of the sylvian fissure can be used to determine whether the brain tissue of the person to be tested is atrophied, so that the width of the sylvian fissure is increased, thereby obtaining the probability of brain CT Alzheimer's disease discrimination, so as to objectively and accurately describe the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease. When determining the image similarity between the brain CT image of the person to be tested and the brain CT image of Parkinson's disease, the second critical point can be screened by the gradient change of the pixel values ​​of the pixel points on the second circular test line, and the average width of the substantia nigra of the midbrain can be determined by the conditional function, and then the average width of the substantia nigra of the midbrain of the person to be tested is determined whether the substantia nigra of the midbrain of the person to be tested is atrophied, thereby obtaining the image similarity between the brain CT image of the person to be tested and the brain CT image of Parkinson's disease, so as to accurately and objectively describe the possibility of suffering from Alzheimer's disease. When determining the image similarity between the brain CT image of the person to be tested and the brain CT image of hydrocephalus, the ratio of the radius of the second circumscribed circle to the radius of the first circumscribed circle can be used to describe whether the ventricles are significantly enlarged, thereby obtaining the image similarity between the brain CT image of the person to be tested and the brain CT image of hydrocephalus, so as to objectively and accurately describe the possibility that the person to be tested suffers from hydrocephalus.

[0020] It should be understood that the above general description and the following detailed description are exemplary and explanatory only and do not limit the present invention. Other features and aspects of the present invention will become more apparent from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other embodiments can be obtained based on these drawings without creative work.

[0022] Figure 1 A schematic diagram of a brain disease prevention information push system according to an embodiment of the present invention is exemplarily shown. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0025] Figure 1 A schematic diagram of a brain disease prevention information push system according to an embodiment of the present invention is exemplarily shown, wherein the system comprises:

[0026] Motion detection equipment, communication components, processors, cloud servers and information receiving terminals;

[0027] The motion detection device includes a waist wearable motion detection device, a knee wearable motion detection device, and an ankle wearable motion detection device;

[0028] The processor is used to:

[0029] When the person to be tested wears the motion detection device, the relative position relationship between the multiple motion detection devices is obtained at multiple times within a preset time period;

[0030] Determining the posture data of the person to be measured according to the relative position relationship;

[0031] Sending the posture data at multiple times within a preset time period to a cloud server through a communication component;

[0032] The cloud server is used for:

[0033] Receiving posture data of the person to be tested;

[0034] Retrieving the brain CT image and EEG data of the person to be tested from the memory;

[0035] Obtaining brain disease prevention information of the person to be tested according to the posture data, the brain CT image and the EEG data;

[0036] The brain disease prevention information is pushed to an information receiving terminal.

[0037] According to the brain disease prevention information push system of an embodiment of the present invention, the posture data of the person to be tested can be monitored in real time by wearing motion detection equipment, providing a data basis for timely detection of gait abnormalities, and can generate brain disease prevention information together with CT images, EEG data and other information, so as to timely detect early brain diseases and push them to the information receiving terminal to prompt the person to be tested to pay attention to brain diseases.

[0038] According to one embodiment of the present invention, the waist-wearable motion detection device includes a waist brace and a first ultrasonic sensor integrated in the waist brace; the knee-wearable motion detection device includes a knee brace and a second ultrasonic sensor integrated in the knee brace; the ankle-wearable motion detection device includes an ankle brace and a third ultrasonic sensor integrated in the ankle brace; wherein the frequencies of ultrasonic waves emitted by the first ultrasonic sensor, the second ultrasonic sensor and the third ultrasonic sensor are different from each other, so that each ultrasonic sensor can recognize the echo of the ultrasonic wave emitted by itself.

[0039] According to one embodiment of the present invention, when a person to be tested wears a motion detection device, the relative position relationship between multiple motion detection devices is obtained at multiple moments within a preset time period, including: setting multiple measurement marks at preset positions in a test area; emitting ultrasonic waves with a first frequency to multiple measurement marks through a first ultrasonic sensor, receiving ultrasonic echoes with the first frequency from multiple measurement marks, and determining a first distance between the first ultrasonic sensor and the multiple measurement marks according to a time difference between the ultrasonic waves with the first frequency and the ultrasonic echoes with the first frequency; determining a first coordinate of the waist-worn motion detection device in the test area according to the first distance; emitting ultrasonic waves with a second frequency to multiple measurement marks through a second ultrasonic sensor, receiving ultrasonic echoes with the second frequency from multiple measurement marks wave echo, and determine the second distance between the second ultrasonic sensor and the multiple measurement marks according to the time difference between the ultrasonic wave of the second frequency and the ultrasonic wave echo of the second frequency; determine the second coordinate of the knee wearable motion detection device in the test area according to the second distance; transmit the ultrasonic wave with a third frequency to the multiple measurement marks through the third ultrasonic sensor, receive the ultrasonic wave echo with the third frequency of the multiple measurement marks, and determine the third distance between the third ultrasonic sensor and the multiple measurement marks according to the time difference between the ultrasonic wave of the third frequency and the ultrasonic wave echo of the third frequency; determine the third coordinate of the ankle wearable motion detection device in the test area according to the third distance; determine the relative position relationship between the multiple motion detection devices according to the first coordinate, the second coordinate and the third coordinate.

[0040] According to one embodiment of the present invention, based on the time difference between the reception time of the ultrasonic echo and the emission time of the ultrasonic wave, the distance between the ultrasonic sensor and each measurement mark can be determined, and the position of each measurement mark is known. Therefore, through the analysis of the above ultrasonic echo, the distance between the ultrasonic sensor and multiple known positions can be determined, and the coordinates of the ultrasonic sensor can be obtained by the simultaneous equations of the distance and the coordinates of the known positions. The coordinates of the ultrasonic sensor are the coordinates of the motion detection device. Further, the vectors between the various motion detection devices can be determined based on the coordinates of the motion detection device, and then the relative position relationship between the multiple motion detection devices can be determined. The vectors between the various motion detection devices can also be used as the posture data of the person to be measured. The processor can send the posture data at multiple times to the cloud server for analysis.

[0041] According to one embodiment of the present invention, when the cloud server receives the posture data of the person to be tested, it can also call the brain CT image and EEG data of the person to be tested for joint analysis. According to the posture data, brain CT image and EEG data, the brain disease prevention information of the person to be tested is obtained, including: retrieving reference posture data of multiple brain diseases from the memory; determining the gait discrimination information of the person to be tested according to the posture data and the reference posture data; detecting the brain CT image through the image detection model to obtain the first position area where the brain tissue is located; determining the brain CT discrimination information of the person to be tested according to the first position area; determining the EEG discrimination information of the person to be tested according to the EEG data; determining the brain disease prevention information of the person to be tested according to the gait discrimination information, the brain CT discrimination information and the EEG discrimination information.

[0042] According to one embodiment of the present invention, the memory may include a historical database, which may include various types of historical data of multiple patients. Reference posture data of Alzheimer's patients, reference posture data of Parkinson's patients, and reference posture data of hydrocephalus patients can be obtained from the historical database. Doctors can select multiple representative (for example, patients with long-term illness and severe illness) Alzheimer's disease patients' posture data at multiple moments when walking, and make the change rate of the posture data of multiple Alzheimer's disease patients when walking consistent by slow motion or fast forward, etc. (that is, obtain the posture data of multiple Alzheimer's disease patients with consistent walking speed), and can fuse the posture data of multiple Alzheimer's disease patients when walking (for example, the posture data includes the relative position relationship between each wearable motion detection device, and the relative position relationship can be represented by the position vector between each wearable motion detection device, and the position vectors of the same parts of multiple Alzheimer's disease patients can be averaged for fusion, for example, the position vectors between the wearable motion detection devices at the knees and waists of multiple Alzheimer's disease patients are averaged), to obtain reference posture data of Alzheimer's disease patients. Similarly, reference posture data of Parkinson's disease patients and reference posture data of hydrocephalus patients can be obtained in a similar manner to the reference posture data of Alzheimer's disease patients, which will not be repeated here.

[0043] According to one embodiment of the present invention, the gait discrimination information of the person to be tested is determined based on the posture data and the reference posture data, including: determining the gait similarity between the posture data and the reference posture data of various brain diseases through a posture similarity recognition model; and obtaining the gait discrimination information of the person to be tested based on the gait similarity between the reference posture data and the reference posture data of various brain diseases.

[0044] According to one embodiment of the present invention, the posture data of the person to be tested may be feature-matched with reference posture data of Alzheimer's disease patients, reference posture data of Parkinson's disease patients, and reference posture data of hydrocephalus patients, respectively, so as to determine gait discrimination information. For example, the gait discrimination information is information in the form of a vector, and the vector may have three components, namely, the gait similarity between the posture data of the person to be tested and the reference posture data of Alzheimer's disease patients, the gait similarity between the posture data of the person to be tested and the reference posture data of Parkinson's disease patients, and the gait similarity between the posture data of the person to be tested and the reference posture data of hydrocephalus patients. Therefore, based on the maximum value of the three similarities, it can be determined which brain disease the posture data of the person to be tested has the greatest similarity with, that is, which disease the gait characteristics of the person to be tested are most likely to be.

[0045] According to one embodiment of the present invention, the gait similarity between the reference posture data and the reference posture data of various brain diseases is determined by a posture similarity recognition model, including: inputting the posture data of patients diagnosed with Alzheimer's disease in the historical database, the reference posture data of patients with Alzheimer's disease, the reference posture data of patients with Parkinson's disease, and the reference posture data of patients with hydrocephalus into the posture similarity recognition model to obtain a first discriminant vector; inputting the posture data of patients diagnosed with Parkinson's disease in the historical database, the reference posture data of patients with Alzheimer ... , the reference posture data of patients with Parkinson's disease and the reference posture data of patients with hydrocephalus, are input into the posture similarity recognition model to obtain a second discriminant vector; the posture data of patients with hydrocephalus in the historical database, the reference posture data of patients with Alzheimer's disease, the reference posture data of patients with Parkinson's disease and the reference posture data of patients with hydrocephalus are input into the posture similarity recognition model to obtain a third discriminant vector; the posture data of control persons without brain diseases, the reference posture data of patients with Alzheimer's disease, the reference posture data of patients with Parkinson's disease and the reference posture data of patients with hydrocephalus are input into the posture similarity recognition model to obtain a third discriminant vector. The reference posture data is input into a posture similarity recognition model to obtain a fourth discriminant vector; the posture data of an experimental patient suffering from any one of three brain diseases, namely, Alzheimer's disease, Parkinson's disease and hydrocephalus, the reference posture data of the Alzheimer's disease patient, the reference posture data of the Parkinson's disease patient and the reference posture data of the hydrocephalus patient are input into the posture similarity recognition model to obtain a fifth discriminant vector, wherein the posture data of the experimental patient is not stored in the historical database; according to the first discriminant vector, the second discriminant vector and the third discriminant vector, vector, the fourth discriminant vector and the fifth discriminant vector to obtain the loss function of the posture similarity recognition model; the posture similarity recognition model is trained according to the loss function of the posture similarity recognition model to obtain the trained posture similarity recognition model; the posture data of the person to be tested, the reference posture data of the Alzheimer's disease patient, the reference posture data of the Parkinson's disease patient and the reference posture data of the hydrocephalus patient are input into the trained posture similarity recognition model to obtain the gait similarity between the reference posture data and the reference posture data of various brain diseases.

[0046] According to one embodiment of the present invention, the posture similarity recognition model is a BP neural network model, which is used to determine the feature similarity between the posture data of the person to be tested and the reference posture data of Alzheimer's patients, the reference posture data of Parkinson's patients, and the reference posture data of hydrocephalus patients, and obtain gait discrimination information. Since the disease characteristics of the patient are not obvious in the early stage of the disease, it is difficult to manually judge the similarity between the posture data of the person to be tested and various reference posture data to determine which disease the person is most likely to suffer from. Therefore, the posture similarity recognition model can be used for judgment, so that the characteristics of different dimensions of the posture data of the person to be tested and the similarity of various reference posture data can be analyzed.

[0047] According to one embodiment of the present invention, before using the posture similarity recognition model to process the posture data, the posture similarity recognition model can be trained. The posture data in the historical database, the posture data of control persons without brain diseases, and the posture data of experimental patients who do not belong to the historical database but have brain diseases can be used as training samples, so that the posture similarity recognition model can be trained to distinguish brain diseases and exclude non-disease factors, thereby improving the accuracy of the posture similarity recognition model.

[0048] According to one embodiment of the present invention, the posture data of patients diagnosed with Alzheimer's disease in the historical database and the above multiple reference posture data can be input into the posture similarity recognition model to obtain a first discriminant vector, so as to determine whether the posture similarity recognition model can judge that the posture data of patients diagnosed with Alzheimer's disease has the highest feature similarity with the reference posture data of patients with Alzheimer's disease, and has a lower feature similarity with the other two reference posture data.

[0049] According to one embodiment of the present invention, the posture data of patients diagnosed with Parkinson's disease in the historical database and the above multiple reference posture data can be input into the posture similarity recognition model to obtain a second discriminant vector, so as to determine whether the posture similarity recognition model can judge that the posture data of patients diagnosed with Parkinson's disease has the highest feature similarity with the reference posture data of patients with Parkinson's disease, and has a lower feature similarity with the other two reference posture data.

[0050] According to one embodiment of the present invention, the posture data of patients diagnosed with hydrocephalus in the historical database and the above multiple reference posture data can be input into the posture similarity recognition model to obtain a third discriminant vector, thereby determining whether the posture similarity recognition model can determine that the posture data of patients diagnosed with hydrocephalus has the highest feature similarity with the reference posture data of hydrocephalus patients, and has a lower feature similarity with the other two reference posture data.

[0051] According to one embodiment of the present invention, the postures of patients with certain brain diseases and those without brain diseases are not completely different when walking, and there may be some similarities. These similarities may disturb the posture similarity recognition model's discrimination of patients with brain diseases as non-disease factors, so that the posture similarity recognition model can be trained to exclude the interference of non-disease factors, thereby focusing on the recognition of the characteristics of different diseases. The posture data of control persons without brain diseases and the above multiple reference posture data can be input into the posture similarity recognition model to obtain a fourth discriminant vector, so as to determine whether the posture similarity recognition model can determine that the posture data of control persons without brain diseases and the above multiple reference posture data have low feature similarities, and whether non-disease factors can be excluded.

[0052] According to one embodiment of the present invention, the posture data of experimental patients who do not belong to the historical database but suffer from any one of the three brain diseases of Alzheimer's disease, Parkinson's disease and hydrocephalus, as well as the above multiple reference posture data can be input into the posture similarity recognition model to obtain the fifth discriminant vector to determine whether the posture similarity recognition model can be applied to the discrimination of posture data outside the historical database, thereby improving the robustness of the posture similarity recognition model.

[0053] According to one embodiment of the present invention, according to the first discriminant vector, the second discriminant vector, the third discriminant vector, the fourth discriminant vector and the fifth discriminant vector, a loss function of the posture similarity recognition model is obtained, including: according to formula (1), the loss function of the posture similarity recognition model is obtained: ,

[0054] (1)

[0055] in, is the first discriminant vector corresponding to the posture data of the i-th patient diagnosed with Alzheimer's disease, is the preset first label vector, is the preset second label vector, is the preset third label vector, is the number of pose data of patients diagnosed with Alzheimer's disease in a training batch, is the second discriminant vector corresponding to the posture data of the jth patient diagnosed with Parkinson's disease, is the number of posture data of patients diagnosed with Parkinson's disease in a training batch, is the third discriminant vector corresponding to the posture data of the kth patient diagnosed with hydrocephalus, is the number of posture data of patients diagnosed with hydrocephalus in a training batch, and is the logical operator of "and", is the two-norm, is the conditional function, is the fourth discriminant vector corresponding to the posture data of the t-th control person, is the number of pose data of the control personnel in a training batch, for The transposed vector of is the fifth discriminant vector corresponding to the posture data of the sth experimental patient, is the number of experimental patient pose data in a training batch, is the label vector of the sth experimental patient, , , , and is the preset weight, i≤ , j≤ , k≤ , t≤ , s≤ , and i, 、j、 , k, ,t, , s and All are positive integers.

[0056] According to one embodiment of the present invention, the first annotation vector is the theoretical value of the gait discrimination information obtained by the posture similarity recognition model for processing the posture data of the Alzheimer's patient. For example, in the first annotation vector, the posture data has the highest similarity with the reference posture data of the Alzheimer's patient (for example, the similarity is 100%), and the similarity with the other two reference posture data is lower (for example, the similarity is 0). The second annotation vector is the theoretical value of the gait discrimination information obtained by the posture similarity recognition model for processing the posture data of the Parkinson's patient. For example, in the second annotation vector, the posture data has the highest similarity with the reference posture data of the Parkinson's patient (for example, the similarity is 100%), and the similarity with the other two reference posture data is lower (for example, the similarity is 0). The third annotation vector posture similarity recognition model processes the posture data of the hydrocephalus patient and obtains the theoretical value of the gait discrimination information. For example, in the third annotation vector, the posture data has the highest similarity with the reference posture data of the hydrocephalus patient (for example, the similarity is 100%), and has a lower similarity with the other two reference posture data (for example, the similarity is 0).

[0057] According to one embodiment of the present invention, in formula (1), The modulus of the difference between the first discriminant vector and the first labeled vector is smaller than the modulus of the difference between the first discriminant vector and the second labeled vector, and smaller than the modulus of the difference between the first discriminant vector and the third labeled vector, that is, the first discriminant vector has the highest similarity with the first labeled vector. When this condition is met, it means that the judgment of the posture similarity recognition model is correct, and there is only an error with the first labeled vector in the specific feature similarity value. Therefore, the conditional function value can be made the modulus of the difference between the first discriminant vector and the first labeled vector. Otherwise, it means that the judgment of the posture similarity recognition model is wrong, that is, the first discriminant vector itself is an incorrect discriminant vector, which itself is a judgment error, and its modulus value can be used as the conditional function value. Therefore, the conditional function value is the error in the two cases, and the average value of the error corresponding to the posture data of patients diagnosed with Alzheimer's disease in a training batch can be obtained as a loss function item. The loss function can be reduced during the training process, thereby reducing the error and improving the judgment accuracy of the posture data of patients diagnosed with Alzheimer's disease.

[0058] According to one embodiment of the present invention, The modulus of the difference between the second discriminant vector and the second labeled vector is smaller than the modulus of the difference between the second discriminant vector and the first labeled vector, and smaller than the modulus of the difference between the second discriminant vector and the third labeled vector, that is, the second discriminant vector has the highest similarity with the second labeled vector. When this condition is met, it means that the judgment of the posture similarity recognition model is correct, and there is only an error in the specific feature similarity value with the second labeled vector. Therefore, the conditional function value can be made the modulus of the difference between the second discriminant vector and the second labeled vector. Otherwise, it means that the judgment of the posture similarity recognition model is wrong, that is, the second discriminant vector itself is an incorrect discriminant vector, which itself is a judgment error, and its modulus value can be used as the conditional function value. Therefore, the conditional function value is the error in the two cases, and the average value of the error corresponding to the posture data of patients diagnosed with Parkinson's disease in a training batch can be obtained as a loss function item. The loss function can be reduced during the training process, thereby reducing the error and improving the judgment accuracy of the posture data of patients diagnosed with Parkinson's disease.

[0059] According to one embodiment of the present invention, The modulus of the difference between the third discriminant vector and the third labeled vector is smaller than the modulus of the difference between the third discriminant vector and the first labeled vector, and smaller than the modulus of the difference between the third discriminant vector and the second labeled vector, that is, the third discriminant vector has the highest similarity with the third labeled vector. When this condition is met, it means that the judgment of the posture similarity recognition model is correct, and there is only an error in the specific feature similarity value with the third labeled vector. Therefore, the conditional function value can be made the modulus of the difference between the third discriminant vector and the third labeled vector. Otherwise, it means that the judgment of the posture similarity recognition model is wrong, that is, the third discriminant vector itself is an incorrect discriminant vector, which itself is a judgment error, and its modulus value can be used as the conditional function value. Therefore, the conditional function value is the error in the two cases, and the average value of the error corresponding to the posture data of patients diagnosed with hydrocephalus in a training batch can be obtained as a loss function item. The loss function can be reduced during the training process, thereby reducing the error and improving the judgment accuracy of the posture data of patients diagnosed with hydrocephalus.

[0060] According to one embodiment of the present invention, the first annotation vector, the second annotation vector and the third annotation vector are all vectors composed of similarities with three brain diseases, and among the three similarity values, the similarity with the reference posture data of patients with one of the diseases is higher, and the similarity with the other two diseases is lower. Therefore, by adding the first annotation vector, the second annotation vector and the third annotation vector, a vector composed of similarity values ​​with high similarities to the three brain diseases can be obtained, for example, three vectors composed of 100%. It is the cosine similarity between the fourth discriminant vector and the sum of the above three labeled vectors. Since the fourth discriminant vector is the discriminant vector obtained by processing the posture data of the control persons without brain diseases, the fourth discriminant vector theoretically has low similarity with the three brain diseases. Therefore, the similarity with the sum of the three labeled vectors should be low. Therefore, the average value of the cosine similarities corresponding to the fourth discriminant vectors of each control person without brain disease can be used as a term of the loss function, so that in the training process, the cosine similarity can be reduced, and the difference between the fourth discriminant vector and the sum of the three labeled vectors can be increased, so as to eliminate the interference of non-disease factors and improve the judgment accuracy of the posture similarity recognition model.

[0061] According to one embodiment of the present invention, It is the modulus of the difference between the fifth discriminant vector corresponding to the experimental patient's posture data and the corresponding annotation vector. The annotation vector is related to the type of brain disease suffered by the experimental patient, and the numerical value of its component is the same as the first annotation vector, the second annotation vector or the third annotation vector mentioned above, which will not be repeated here. The modulus value is the error value of the fifth discriminant vector. The average value of the error value of the fifth discriminant vector corresponding to each experimental patient can be used as a term of the loss function, thereby reducing the error value during the training process, improving the accuracy of the posture similarity recognition model, and making the posture similarity recognition model suitable for the discrimination of posture data outside the historical database, thereby improving the robustness of the posture similarity recognition model.

[0062] According to one embodiment of the present invention, the above five items can be weighted and summed to obtain the loss function of the posture similarity recognition model, and the loss function can be back-propagated to adjust the parameters of the posture similarity recognition model through the gradient descent method, so as to train the posture similarity recognition model and improve the judgment accuracy and robustness. After multiple trainings, a trained posture similarity recognition model can be obtained. The posture similarity recognition model can accurately identify the posture characteristics of the gait of the person to be tested in the early stage of the disease, judge the type of disease, and obtain the gait discrimination information of the person to be tested.

[0063] In this way, the judgment accuracy of the posture similarity recognition model can be trained through the posture data in the historical database, and when the judgment is wrong, the wrong discriminant vector is used as a term of the loss function, thereby increasing the training intensity and reducing the probability of judgment errors. The posture similarity recognition model can also be trained through the posture data of control persons without brain diseases to eliminate the interference of non-disease factors, thereby improving the anti-interference ability and judgment accuracy of the posture similarity recognition model. Furthermore, the robustness of the posture similarity recognition model can be trained through the posture data of experimental patients with brain diseases who do not belong to the historical database, so that the posture similarity recognition model can be applied to the judgment of posture data outside the historical database, thereby improving the performance of the posture similarity recognition model.

[0064] According to one embodiment of the present invention, after the training of the posture similarity recognition model is completed, the trained posture similarity recognition model can be used to determine the gait similarity between the posture data and the reference posture data of various brain diseases, that is, the gait similarity between the gait of the person to be tested and the gait of patients with various brain diseases, thereby forming gait discrimination information.

[0065] According to one embodiment of the present invention, the image detection model is a deep learning neural network model, for example, a convolutional neural network model, which can monitor the first position area where the brain tissue is located in the brain CT image and determine the edge of the first position area.

[0066] According to one embodiment of the present invention, the brain CT image may be a three-dimensional CT image, which may include multiple two-dimensional CT image slices, and the area where the brain tissue is located may be detected in each two-dimensional CT image slice, that is, the first position area in the multiple brain CT images.

[0067] According to one embodiment of the present invention, brain CT discrimination information of a person to be tested is determined based on the first position area, including: obtaining a first circumscribed circle of a first position area in a plurality of brain CT images; obtaining a midbrain CT image in a plurality of brain CT images; obtaining the center of the first circumscribed circle in the midbrain CT image; obtaining an upper brain CT image in a plurality of brain CT images; obtaining the radius of the first circumscribed circle in the upper brain CT image; identifying an area where the ventricles in the upper brain CT image are located, and determining a second circumscribed circle of the area where the ventricles are located; obtaining the radius of the second circumscribed circle; obtaining the brain CT discrimination information of the person to be tested based on the center of the first circumscribed circle in the midbrain CT image, the radius of the first circumscribed circle in the upper brain CT image, and the radius of the second circumscribed circle.

[0068] According to one embodiment of the present invention, the first circumscribed circle of the first position area in each two-dimensional CT image slice can be determined. Further, according to the center of the first circumscribed circle, the first circular test line and the second circular test line are determined, wherein the first circular test line and the second circular test line both have the center of the first circumscribed circle as the center, and the radii of the first circular test line and the second circular test line are both smaller than the radius of the first circumscribed circle, the length ratio of the radius of the first circular test line to the radius of the first circumscribed circle is equal to a first preset ratio, the length ratio of the radius of the second circular test line to the radius of the first circumscribed circle is equal to a second preset ratio, and the first preset ratio is greater than the second preset ratio.

[0069] According to one embodiment of the present invention, the scanning direction of the brain CT image can be from top to bottom, and a two-dimensional CT image slice with a specific serial number can be selected in the middle area in the up-down direction as the midbrain CT image. In the midbrain CT image, the midbrain substantia nigra is included, which is located at the center of the first position area where the brain tissue is located. The center of the first circumscribed circle is determined in the midbrain CT image, and the first circular test line and the second circular test line are determined with the center of the first circumscribed circle as the center. The radius of the first circular test line is greater than the radius of the second circular test line, and both radii are smaller than the radius of the first circumscribed circle. The first circular test line is used to test whether the area at the edge of the brain tissue is atrophied, and whether the lateral fissure between brain tissues in different partitions is widened. The second circular test line is used to test whether the midbrain substantia nigra is degenerated.

[0070] According to one embodiment of the present invention, the upper brain CT image is a two-dimensional CT image slice with a specific serial number in the upper region in the up-down direction. In the upper brain CT image, the region where the ventricles are located is included, which can be identified, for example, by an image recognition model, which is a convolutional neural network model. The second circumscribed circle of the region where the ventricles are located can be determined, and the radius of the second circumscribed circle can be determined. The radius of the first circumscribed circle in the upper brain CT image can also be determined. The radius of the second circumscribed circle and the radius of the first circumscribed circle can be used to determine whether the region where the ventricles are located is enlarged.

[0071] According to one embodiment of the present invention, brain CT discrimination information of a person to be tested is obtained based on the center of a first circumscribed circle in a midbrain CT image, the radius of the first circumscribed circle in an upper brain CT image, and the radius of a second circumscribed circle, including: determining the image similarity between the brain CT image of the person to be tested and brain CT images of multiple diseases based on the center of the first circumscribed circle in the midbrain CT image, the radius of the first circumscribed circle in the upper brain CT image, and the radius of the second circumscribed circle; and determining the brain CT discrimination information of the person to be tested based on the image similarity.

[0072] According to one embodiment of the present invention, the image similarity between the brain CT image of the person to be tested and the brain CT images of multiple diseases is determined according to the center of the first circumscribed circle in the midbrain CT image, the radius of the first circumscribed circle in the upper brain CT image, and the radius of the second circumscribed circle, including: determining the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease according to the first circular test line in the midbrain CT image; determining the image similarity between the brain CT image of the person to be tested and the brain CT image of Parkinson's disease according to the second circular test line in the midbrain CT image; determining the image similarity between the brain CT image of the person to be tested and the brain CT image of hydrocephalus according to the radius of the first circumscribed circle in the upper brain CT image and the radius of the second circumscribed circle.

[0073] According to one embodiment of the present invention, based on the first circular test line in the midbrain CT image, determining the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease includes: obtaining the first critical point selection condition on the first circular test line according to formula (2): ,

[0074] (2)

[0075] in, is the coordinate of the ath pixel point on the first circular test line, is the pixel value of the ath pixel on the first circular test line, is the direction vector at the ath pixel point on the first circular test line, The pixel value of the a-th pixel on the first circular test line is along the direction vector The gradient of is the first preset ratio; meets the critical point selection condition on the first circular test line The pixel points constitute the first critical point set on the first circular test line; according to formula (3), the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease is obtained: ,

[0076] (3)

[0077] in, is the average value of the pixel values ​​of multiple pixels between the bth first critical point and the b+1th first critical point on the first circular test line, is the first preset pixel value, is the sequence number of the bth first critical point among the multiple pixel points on the first circular test line, is the sequence number of the b+1th first critical point among the multiple pixel points on the first circular test line, is the first preset width, , is the number of first critical points in the first critical point set, is the conditional function, b≤ , b. and All are positive integers.

[0078] According to one embodiment of the present invention, in formula (2), The pixel value of the a-th pixel on the first circular test line is along the direction vector The modulus of the gradient represents the rate of change of the pixel value along the direction vector. If the rate of change is greater than or equal to the first preset ratio, it means that the first circular test line crosses the boundary between the brain tissue and the background area along the direction of the direction vector. Therefore, the a-th pixel point is the first critical point on the boundary between the brain tissue and the background area. All the first critical points on the first circular test line can be determined, and the first critical point set can be obtained.

[0079] According to one embodiment of the present invention, the typical features of brain CT of Alzheimer's disease are atrophy of the edge of the brain tissue and widening of the lateral fissure between different partitions of brain tissue. If the first circular test line between the two first critical points is located in the background area, the first circular test line between the two first critical points crosses the lateral fissure between the different partitions of brain tissue. Therefore, if the distance between the two first critical points is increased compared to the normal level, it can be used as a precursor of Alzheimer's disease.

[0080] According to one embodiment of the present invention, in formula (3), Indicated in If the condition is true, the outer condition function value is 1, otherwise .

[0081] According to one embodiment of the present invention, Indicated in If it holds, the inner condition function value is , otherwise 0. The average value of the pixel values ​​of the multiple pixels between the bth first critical point and the b+1th first critical point on the first circular test line is higher than or equal to the first preset pixel value, indicating that the pixel values ​​of the pixels between the two first critical points are closer to the background color. Therefore, the first circular test line between the two first critical points passes through the lateral fissure between brain tissues of different partitions. In this case, the inner conditional function value is , which is the serial number difference between the two first critical points, and can represent the distance between the two first critical points on the first circular test line. If the above conditions do not hold, the inner condition function value is 0. It is the sum of the distances to the first critical points at both ends of the lateral fissure, and can also be expressed as the sum of the widths of the lateral fissure. ,in, It means that the first circular test line between the two first critical points crosses the side crack, then the conditional function value is 1, otherwise it is 0, so, is the number of lateral fissures. It can represent the average width of the lateral fissure. If the average width is greater than or equal to the first preset width, the outer conditional function value is 1, indicating that the probability that the person to be tested suffers from Alzheimer's disease is 1, and the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease is also 1.

[0082] According to one embodiment of the present invention, if the average width is less than the first preset width, the outer condition function value is , which is the ratio of the average width of the sylvian fissure to the first preset width. In this case, the larger the average width of the sylvian fissure, the greater the possibility that the person to be tested suffers from Alzheimer's disease. This value can be used as the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease.

[0083] According to one embodiment of the present invention, the value of the outer condition function may be used as the brain CT Alzheimer's disease discrimination probability to describe the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease.

[0084] In this way, the first critical point can be screened by the gradient change of the pixel values ​​of the pixel points on the first circular test line, and the average width of the sylvian fissure can be determined by the conditional function, and then the average width of the sylvian fissure can be used to determine whether the brain tissue of the person to be tested is atrophied, thereby increasing the width of the sylvian fissure, and thus obtaining the probability of Alzheimer's disease discrimination in brain CT, so as to objectively and accurately describe the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease.

[0085] According to one embodiment of the present invention, based on the second circular test line in the midbrain CT image, determining the image similarity between the brain CT image of the person to be tested and the brain CT image of Parkinson's disease includes: obtaining the second critical point selection condition on the second circular test line according to formula (4): ,

[0086] (4)

[0087] in, is the coordinate of the cth pixel point on the second circular test line, is the pixel value of the cth pixel on the second circular test line, is the direction vector at the c-th pixel point on the second circular test line, is the pixel value of the cth pixel point on the second circular test line along the direction vector The gradient of is the second preset ratio; and meets the second critical point selection condition on the second circular test line The pixel points constitute the second critical point set on the second circular test line; according to formula (5), the image similarity between the brain CT image of the person to be tested and the brain CT image of Parkinson's disease is determined ,

[0088] (5)

[0089] in, is the average value of the pixel values ​​of multiple pixel points between the dth second critical point and the d+1th second critical point on the second circular test line, is the second preset pixel value, is the serial number of the d-th second critical point among the multiple pixel points on the second circular test line, is the serial number of the d+1th second critical point among the multiple pixel points on the second circular test line, is the second preset width, , is the number of second critical points in the second critical point set, is the conditional function, d, and All are positive integers.

[0090] According to one embodiment of the present invention, the pixel values ​​of the substantia nigra in the midbrain are significantly different from those of other brain tissues. In addition, in the case of Parkinson's disease, due to damage to dopamine neurons, the substantia nigra in the midbrain atrophies and becomes narrower than normal. Therefore, whether there is a precursor to Parkinson's disease can be determined based on the width of the substantia nigra in the midbrain.

[0091] According to one embodiment of the present invention, similar to the screening method of the first critical point, the second critical point between the substantia nigra and other brain tissues on the second circular test line can be screened by selecting the second critical point based on the gradient to form a second critical point set.

[0092] According to one embodiment of the present invention, in formula (5), is the total width of multiple midbrain substantia nigra. is the number of substantia nigra in the midbrain, therefore, The outer conditional function indicates that if the average width of the substantia nigra is less than or equal to the second preset width, it means that the average width of the substantia nigra is too low, and the person to be tested has a precursor of Parkinson's disease, and the value of the outer conditional function is 1; otherwise, the value of the outer conditional function is , that is, the ratio between the second preset width and the average width of the substantia nigra in the midbrain. The larger the average width of the substantia nigra in the midbrain, the lower the ratio, and the lower the probability of suffering from Parkinson's disease. The outer condition function value in formula (5) can be used as the image similarity between the brain CT image of the person to be tested and the brain CT image of Parkinson's disease.

[0093] In this way, the second critical point can be screened by the gradient change of the pixel values ​​of the pixel points on the second circular test line, and the average width of the substantia nigra of the midbrain can be determined by the conditional function. Then, whether the substantia nigra of the midbrain of the person to be tested is atrophied can be determined by the average width of the substantia nigra of the midbrain, thereby obtaining the image similarity between the brain CT image of the person to be tested and the brain CT image of Parkinson's disease, so as to accurately and objectively describe the possibility of suffering from Alzheimer's disease.

[0094] According to one embodiment of the present invention, the image similarity between the brain CT image of the person to be tested and the brain CT image of hydrocephalus is determined according to the radius of the first circumscribed circle and the radius of the second circumscribed circle in the upper brain CT image, including: determining the image similarity between the brain CT image of the person to be tested and the brain CT image of hydrocephalus according to formula (6): ,

[0095] (6)

[0096] in, is the radius of the second circumscribed circle, is the radius of the first circumscribed circle, is the third preset ratio, is the conditional function.

[0097] According to one embodiment of the present invention, in formula (6), the conditional function may represent that when the ratio of the radius of the second circumscribed circle to the radius of the first circumscribed circle is greater than or equal to the third preset ratio, the conditional function value is 1, otherwise it is If the ratio is greater than or equal to the third preset ratio, it means that the ventricle area is significantly enlarged, there is a precursor to hydrocephalus, and the conditional function value is 1. If the ratio is less than the third preset ratio, the ratio of the ratio to the third preset ratio can be used as the conditional function value, indicating that the larger the ratio, the greater the possibility of hydrocephalus. The conditional function value can be used as the image similarity between the brain CT image of the person to be tested and the brain CT image of hydrocephalus to describe the possibility that the person to be tested suffers from hydrocephalus.

[0098] In this way, the ratio of the radius of the second circumscribed circle to the radius of the first circumscribed circle can be used to describe whether the ventricles are significantly enlarged, thereby obtaining the image similarity between the brain CT image of the person to be tested and the brain CT image of hydrocephalus, so as to objectively and accurately describe the possibility that the person to be tested suffers from hydrocephalus.

[0099] According to one embodiment of the present invention, the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease, the image similarity between the brain CT image of the person to be tested and the brain CT image of Parkinson's disease, and the image similarity between the brain CT image of the person to be tested and the brain CT image of hydrocephalus can be used as components of brain CT discrimination information in vector form, thereby constituting brain CT discrimination information.

[0100] According to one embodiment of the present invention, the EEG data of the person to be tested can be obtained, and the EEG discrimination information of the person to be tested can be obtained based on the EEG data. The EEG discrimination information of the person to be tested is determined based on the EEG data, including: determining the waveform feature similarity between the EEG data of the person to be tested and the EEG data of multiple brain diseases; determining the EEG discrimination information of the person to be tested based on the waveform feature similarity.

[0101] According to one embodiment of the present invention, the characteristics of the EEG data of patients with Alzheimer's disease include slowing of the alpha rhythm in the occipital region, a significant increase in diffuse slow wave activity, and a decrease in fast wave activity in the EEG. The intensity of low-frequency brain waves on the EEG of patients with Parkinson's disease is weak, and the reaction time for the EEG potential to reach the peak is long. The EEG of patients with hydrocephalus may have characteristics such as overall slowing of the rhythm. Therefore, the EEG data of different patients in the historical database can be obtained, and the frequency domain data of the EEG data can be obtained by fast Fourier transform, and the extreme points in the frequency domain data can be used as eigenvalues ​​to form feature vectors of different types of brain diseases. Subsequently, the EEG data of the person to be tested can be fast Fourier transformed to obtain frequency domain data, and then the extreme points in the frequency domain data can be used as eigenvalues ​​to obtain the feature vector of the EEG data of the person to be tested. Further, the waveform feature similarity of the EEG data of the person to be tested and the EEG data of various brain diseases can be determined by comparing the similarity between the feature vector of the EEG data of the person to be tested and the feature vector of various brain diseases. In another example, the difference between the frequency domain data of the EEG data of different patients in the database and the frequency domain data of the EEG data of the person to be tested at different frequencies can also be solved, and then the difference at different frequencies can be integrated to obtain the difference between the EEG data of the person to be tested and the EEG data of patients with various types of brain diseases, and then the relative difference between the EEG data of the person to be tested and the EEG data of patients with various types of brain diseases can be solved, and the waveform feature similarity between the EEG data of the person to be tested and the EEG data of patients with various types of brain diseases can be obtained by subtracting the relative difference from 1. Furthermore, the waveform feature similarity between the EEG data of the person to be tested and the EEG data of patients with various brain diseases can be used to form EEG discrimination information.

[0102] According to one embodiment of the present invention, the gait discrimination information, the brain CT discrimination information and the EEG discrimination information can be weighted and summed, and in the obtained sum vector, each component can represent the similarity between the person to be tested and the early symptoms of various brain diseases. Brain disease prevention information can be generated based on the sum vector, and in the brain disease prevention information, the similarity between the person to be tested and the early symptoms of various brain diseases can be described respectively. Furthermore, the brain disease prevention information can be pushed to an information receiving terminal, such as the mobile phone of the person to be tested, the doctor's computer, etc., so as to prompt the person to be tested to prevent the brain disease from further deepening.

[0103] According to the brain disease prevention information push system of the embodiment of the present invention, the posture data of the person to be tested can be monitored in real time by wearing a motion detection device, providing a data basis for timely detection of gait abnormalities, and can generate brain disease prevention information together with CT images, EEG data and other information, which can timely detect early brain diseases and push them to the information receiving terminal to prompt the person to be tested to pay attention to brain diseases. When training the posture similarity recognition model, the judgment accuracy of the posture similarity recognition model can be trained by the posture data in the historical database, and when the judgment is wrong, the wrong discriminant vector is used as a loss function, thereby improving the training intensity and reducing the probability of judgment errors. The posture data of the control person without brain disease can also be used to train the posture similarity recognition model to eliminate the interference of non-disease factors, thereby improving the anti-interference ability and judgment accuracy of the posture similarity recognition model. Further, the robustness of the posture similarity recognition model can be trained by the posture data of experimental patients with brain diseases who do not belong to the historical database, so that the posture similarity recognition model can be applied to the discrimination of posture data outside the historical database, and the performance of the posture similarity recognition model can be improved. When determining the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease, the first critical point can be screened by the gradient change of the pixel values ​​of the pixel points on the first circular test line, and the average width of the sylvian fissure can be determined by the conditional function, and then the average width of the sylvian fissure can be used to determine whether the brain tissue of the person to be tested is atrophied, so that the width of the sylvian fissure is increased, thereby obtaining the probability of brain CT Alzheimer's disease discrimination, so as to objectively and accurately describe the image similarity between the brain CT image of the person to be tested and the brain CT image of Alzheimer's disease. When determining the image similarity between the brain CT image of the person to be tested and the brain CT image of Parkinson's disease, the second critical point can be screened by the gradient change of the pixel values ​​of the pixel points on the second circular test line, and the average width of the substantia nigra of the midbrain can be determined by the conditional function, and then the average width of the substantia nigra of the midbrain of the person to be tested is determined whether the substantia nigra of the midbrain of the person to be tested is atrophied, thereby obtaining the image similarity between the brain CT image of the person to be tested and the brain CT image of Parkinson's disease, so as to accurately and objectively describe the possibility of suffering from Alzheimer's disease. When determining the image similarity between the brain CT image of the person to be tested and the brain CT image of hydrocephalus, the ratio of the radius of the second circumscribed circle to the radius of the first circumscribed circle can be used to describe whether the ventricles are significantly enlarged, thereby obtaining the image similarity between the brain CT image of the person to be tested and the brain CT image of hydrocephalus, so as to objectively and accurately describe the possibility that the person to be tested suffers from hydrocephalus.

[0104] It should be understood by those skilled in the art that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments, and the embodiments of the present invention may be deformed or modified in any way without departing from the principles.

[0105] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A brain disease prevention information push system, characterized in that: include: Motion detection equipment, communication components, processors, cloud servers and information receiving terminals; The motion detection device includes a waist wearable motion detection device, a knee wearable motion detection device, and an ankle wearable motion detection device; The processor is used to: When the person to be tested wears the motion detection device, the relative position relationship between the multiple motion detection devices is obtained at multiple times within a preset time period; Determining the posture data of the person to be measured according to the relative position relationship; Sending the posture data at multiple times within a preset time period to a cloud server through a communication component; The cloud server is used for: Receiving posture data of the person to be tested; Retrieving the brain CT image and EEG data of the person to be tested from the memory; Obtaining brain disease prevention information of the person to be tested according to the posture data, the brain CT image and the EEG data; Pushing the brain disease prevention information to an information receiving terminal; According to the posture data, the brain CT image and the EEG data, brain disease prevention information of the person to be tested is obtained, including: Retrieving reference posture data of various brain diseases from a memory; Determining gait discrimination information of the person to be tested according to the posture data and the reference posture data; Using the image detection model, the brain CT image is detected to obtain a first location area where the brain tissue is located; Determining brain CT identification information of the person to be tested according to the first position area; Determining the EEG discrimination information of the person to be tested according to the EEG data; Determining brain disease prevention information of the person to be tested according to the gait discrimination information, the brain CT discrimination information and the EEG discrimination information; Determining brain CT identification information of the person to be tested according to the first position area includes: Acquire a first circumscribed circle of a first position area in a plurality of brain CT images; Among the plurality of brain CT images, a midbrain CT image is acquired; Obtaining the center of the first circumscribed circle in the midbrain CT image; Among the plurality of brain CT images, an upper brain CT image is acquired; Obtaining the radius of the first circumscribed circle in the upper brain CT image; Identify the region where the ventricles are located in the upper brain CT image, and determine a second circumscribed circle of the region where the ventricles are located; Get the radius of the second circumscribed circle; According to the center of the first circumscribed circle in the midbrain CT image, the radius of the first circumscribed circle in the upper brain CT image, and the radius of the second circumscribed circle, the brain CT identification information of the person to be tested is obtained.

2. The brain disease prevention information push system according to claim 1, characterized in that: Determining gait discrimination information of the person to be tested according to the posture data and the reference posture data, including: Determining the gait similarity between the posture data and reference posture data of various brain diseases through a posture similarity recognition model; Gait discrimination information of the person to be tested is obtained according to the gait similarity between the posture data and the reference posture data of various brain diseases.

3. The brain disease prevention information push system according to claim 1, characterized in that: According to the center of the first circumscribed circle in the midbrain CT image, the radius of the first circumscribed circle in the upper brain CT image, and the radius of the second circumscribed circle, the brain CT discrimination information of the person to be tested is obtained, including: Determining the image similarity between the brain CT image of the person to be tested and the brain CT images of multiple diseases according to the center of the first circumscribed circle in the midbrain CT image, the radius of the first circumscribed circle in the upper brain CT image, and the radius of the second circumscribed circle; According to the image similarity, the brain CT identification information of the person to be tested is determined.

4. The brain disease prevention information push system according to claim 1, characterized in that: Determining the EEG discrimination information of the person to be tested according to the EEG data includes: Determine the similarity of waveform characteristics between the EEG data of the person to be tested and the EEG data of various brain diseases; The EEG discrimination information of the person to be tested is determined according to the waveform feature similarity.

5. The brain disease prevention information push system according to claim 1, characterized in that: The waist wearable motion detection device includes a waist protector and a first ultrasonic sensor integrated in the waist protector; the knee wearable motion detection device includes a knee protector and a second ultrasonic sensor integrated in the knee protector; the ankle wearable motion detection device includes an ankle protector and a third ultrasonic sensor integrated in the ankle protector; The frequencies of the ultrasonic waves emitted by the first ultrasonic sensor, the second ultrasonic sensor and the third ultrasonic sensor are different from each other.

6. The brain disease prevention information push system according to claim 5, characterized in that: When the person to be tested wears a motion detection device, the relative position relationship between the multiple motion detection devices is obtained at multiple moments within a preset time period, including: Setting a plurality of measurement marks at preset locations in the test area; Transmitting an ultrasonic wave having a first frequency to a plurality of measurement marks through a first ultrasonic sensor, receiving ultrasonic echoes having the first frequency from the plurality of measurement marks, and determining a first distance between the first ultrasonic sensor and the plurality of measurement marks according to a time difference between the ultrasonic wave having the first frequency and the ultrasonic echoes having the first frequency; Determine a first coordinate of the waist-worn motion detection device in a test area according to the first distance; Transmitting an ultrasonic wave having a second frequency to a plurality of measurement marks through a second ultrasonic sensor, receiving ultrasonic echoes having the second frequency from the plurality of measurement marks, and determining a second distance between the second ultrasonic sensor and the plurality of measurement marks according to a time difference between the ultrasonic wave having the second frequency and the ultrasonic echoes having the second frequency; determining a second coordinate of the knee-wearable motion detection device in the test area according to the second distance; Transmitting an ultrasonic wave having a third frequency to a plurality of measurement marks through a third ultrasonic sensor, receiving ultrasonic echoes having the third frequency from the plurality of measurement marks, and determining a third distance between the third ultrasonic sensor and the plurality of measurement marks according to a time difference between the ultrasonic wave having the third frequency and the ultrasonic echoes having the third frequency; Determining a third coordinate of the ankle-wearable motion detection device in the test area according to the third distance; The relative position relationship between the plurality of motion detection devices is determined according to the first coordinate, the second coordinate and the third coordinate.

Images